Adaptive Lane Marking Recognition During Vehicle Lane Changes
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing travel lane marking recognition systems face difficulties in accurately recognizing lane markings during lane changes due to the changing positions of markings in the image, making early recognition challenging.
Innovation Solution
A travel lane marking recognition apparatus that includes an imaging unit, edge point extractor, candidate detector, probability calculator, recognizer, lane change detector, and condition changer, which adjusts calculation conditions or threshold values during lane changes to enhance the recognition of lane markings by increasing the probability of detection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If the own vehicle is in the midst of changing lanes, then the positions of travel lane markings in the image change significantly over time, but this makes early recognition of travel lane markings difficult
Solution Approach 1:
The patent applies dynamics by making the probability calculation conditions and threshold values adaptive rather than fixed. The system dynamically adjusts the threshold for recognizing travel lane markings based on whether the vehicle is in the midst of a lane change, allowing the recognition system to respond to changing conditions. This enables early recognition during lane changes by lowering thresholds when needed while maintaining high accuracy during normal driving.
Solution Approach 2:
The patent changes parameters (probability calculation conditions and threshold values) based on the vehicle's operational state. When a lane change is detected, the system modifies the threshold parameter to be more permissive, enabling earlier recognition of lane markings that appear in peripheral portions of the image. This parameter adaptation resolves the contradiction by allowing flexible adjustment between speed and accuracy based on contextual conditions.
2Loss of time
If the travel lane marking is in the peripheral portion of the image during lane change, then the marking appears later in the image sequence, but early recognition becomes difficult due to position changes
Solution Approach 1:
The system performs preliminary detection of lane change states and proactively adjusts recognition parameters before the lane marking fully enters the central field of view. By detecting the lane change condition in advance and pre-adjusting the probability thresholds, the system enables earlier recognition of peripheral markings without sacrificing reliability, as the parameter adjustment is based on the detected lane change state rather than waiting for certain visual confirmation.
Solution Approach 2:
The system uses feedback from the lane change detection unit to continuously adjust the probability calculation conditions. When the feedback indicates a lane change is in progress, the system modifies the recognition thresholds to accommodate peripheral markings. This closed-loop feedback mechanism ensures that recognition timing is optimized based on real-time vehicle state information while maintaining detection reliability through continuous monitoring and adjustment.
Data Source
AI summary
A travel lane marking probability calculating unit calculates a travel lane marking probability of each of travel lane marking candidates based on a calculation condition. A travel lane marking recognizing unit recognizes, as a travel lane marking, a travel lane marking candidate having a travel lane marking probability that is a threshold value or higher, among the travel lane marking candidates. A lane change detecting unit detects that an own vehicle is in the midst of a lane change. When the own vehicle is in the midst of a lane change, a condition changing unit changes the calculation condition to allow the travel lane marking probability to be more easily increased compared to when the own vehicle is not in the midst of a lane change, or changes the threshold value to be lower than that when the own vehicle is not in the midst of a lane change.